AWS re:Invent 2022 happens this week. I assume you know what kind of event that is, so I'll move straight to the point. No, I'm not going to write about my predictions, but I'll focus on what I'd wish for instead. I hope the difference between those two is clear:

  • Predictions are about what you can realistically expect (based on the AWS strategy, market, etc.), regardless of whether you'd like it or not.
  • Wishes are about what you'd personally like to get, as you think it would be the most beneficial for you.

My case is rather specific because I am not an AWS customer but an employee. As my wishful thinking can't be about what would benefit my own business most, I'll try to wear an independent developer's hat, looking primarily through the Developer Experience (DevEx) lens.

Disclaimer #1: Obviously, all the opinions and ideas below are my own and should not be associated with the strategy and messaging of Amazon Web Services. I've created this article in my free time, not as part of my daily job.

Disclaimer #2: Many items from my wishlist ... would never make it to AWS re:Invent announcement list. Why so? AWS likes early customer feedback - that's why many services are initially released with limited functionality (MVP style) that gets consecutively expanded over the years. Examples? CDK, calculator.aws, CDK8s, CW Evidently, Fault Injection Simulator, AppRunner. And I have a full backlog of further development ideas for such services - they are not flashy and would not impress many, but they could matter a lot for developers' QoL. I'll skip these.

Great, we should be on the same page now, so let's start daydreaming then. What would I love to see announced during AWS re:Invent 2022?

  1. It's the highest time for NVMe-era, built-from-scratch innovation in the OLTP DB space. It has to support SQL, be truly serverless (compute-wise), and should have new constructs that enable wild-numbers out-scalability (e.g., built-in sharding and proxying). No vacuums, no manual indexing - DB-as-an-endpoint kind of service.
  2. Glue is a powerful service, but sometimes, it reminds me of a cannon used to kill flies. I'd love a simple, SQL-powered ETL service (with simple-to-use, visual processing orchestration) that isn't just serverless but also has a minimum warmup time for cases that do not require spinning a cluster ...
  3. I'd love a scalable, flexible option for CI/CD based on the best industry standards and (preferably) an OSS project (or more, like Tekton and Jenkins X) with a strong community. A managed, serverless service (needless to say).
  4. AWS Resource Explorer has just been released, but I have an appetite for (far) more - a strong bet on "application"/"workload" as a grouping abstraction: not just searching/browsing, but also cost management, visualization (yes!), or clean-up.
  5. I like the idea of using AWS Marketplace as a "control plane" to offer SaaS services to client companies, but the current model (AMIs, Docker containers, or CF templates) seems overly simplified. Even with custom metering, one has to implement too many generic concerns on their own (e.g., securing IP, upgrades, managing state). Having something that more closely resembles a "SaaS control plane as a service" could be an exciting idea.
  6. AWS provides tools to build massive data lakes, but when it comes to governance, their main focus seems to be security. I'm far from saying it's not important, but I'd like to see data lineage-focused tools as well. W/o them, data lakes are much harder to control (data duplication, orchestration validation/maintenance, corrections, etc.).
  7. Last but not least - as everyone knows, DevEx is king, so every reasonable investment here will likely pay off nicely in the future. AWS does a lot when it comes to SDKs, CLIs, and their own web-based "Studios". However, I'd love to see an announcement of a tight collaboration with at least one top-notch IDE vendor, e.g., JetBrains. Yes, I know AWS has got toolkits for all the popular IDEs, but do they stand out in the crowd? Technically, some successful plugins evolved into standalone IDEs over time, slowly becoming default tools of choice for the wider community.

That's my (very subjective) list. What would be on yours? Is there anything particular you'd like to see announced during AWS re:Invent? Something that would change the game for you? Feel free to share it in the comments below - thanks in advance!

Share this post